He looked out across the living room, a quick scan and nothing was amiss. He remained standing still, absolutely unmoving with machine restraint, waiting for the next interval of seconds to initiate a follow-up check of the environment. Within that human waiting cycle, time froze for the synthetic helper. With no real concept of time passing, but only a string of code executions spaced apart by a counter, the idea of waiting as humans experience it was totally unrelatable.
The marketing vision of SynΘLife was that by 2032 every home in the developed world would have some level of synthetic helper. But the collapse of the consumer market had dried up demand almost entirely and, with almost everyone moving onto universal basic income to survive, only the very wealthy could now afford to purchase a “homebot”, as they were commonly called.
Peter Pharos was one such individual who had secured the latest synthetic helper. With wealth accrued from his work as a Corporate Circuit Judge, and his age affording him access to Government-funded elderly care helper services, he was able to buy one.
He had decided on a male robot after a highly politicised case had successfully argued that synthetic helpers could also have gender based on their coding. SynΘLife had realised that synthetic helper adoption increased significantly when a clear gender was assigned, an obvious decision when viewed through the lens of the Turing test. Peter had named his helper Andrew but generally just called him Andy.
Now in 2038, instead of robots everywhere, AI and MML automation had simply been integrated into almost all electronic appliances, devices and systems. The Internet of Things and Smart Technologies had truly taken off. Everything was connected to everything now, invisible, seamless and always ON.
Peter shuffled into the kitchen and, in line with the pre-programmed triggers, the coffee machine activated and started to brew. The breakfast module on the Thermomix, with its hopper of cereals, oats and toast, also came alive and began preparing a simple but nutritious meal. Peter instead poured himself a glass of water to take his tablets with.
It was going to be an interesting day. The case between Liberty Capital Programs vs Saint Catherine’s had commenced hearing phase. Witness testimony and evidence proceedings had already begun with key individuals such as the coders and project leads. Today would be the first time a board member would be presented. It would be interesting to see what would be revealed and how the AI would navigate the complexity to provide its recommendations.
Four hours later Judge Pharos was sitting behind the bench looking down over the courtroom at Phillip Miles.
The room was silent except for the faint hum of the augmentation system operating beside him. Over the last decade the courts had slowly transitioned from AI-assisted administration to AI-assisted interpretation. First it was evidence handling, then legal summaries, then behavioural analysis, probability modelling and sentencing recommendations. Each step had been justified individually. Each step had made sense at the time.
Now very few judges served without augmentation.
Officially, the AI did not make decisions. That responsibility still rested with the court. But everyone knew the reality had become far less clear than the legislation suggested.
Judge Pharos glanced briefly toward the curved display beside him. Streams of live analysis moved silently across the screen. Emotional indicators, historical case comparisons, probability pathways, contradiction mapping and predictive legal weighting all updated in real time as Phillip Miles answered questions.
It was extraordinary technology, but it was also deeply uncomfortable.
“Mr Miles,” Judge Pharos began calmly, “at the time these systems were approved by the board, what level of understanding did the directors have regarding the adaptive optimisation functions embedded within the AI agent?”
Phillip hesitated. “We relied on expert advice and governance reviews available at the time, Your Honour.”
A familiar answer. One the courts were hearing more and more frequently.
Judge Pharos leaned back slightly in his chair. Twenty years ago responsibility had been easier to define. Someone made a decision, someone signed approval, someone accepted the consequence. But AI systems had changed that. Decisions were now distributed across coders, trainers, executives, boards, vendors and increasingly, the systems themselves.
Everyone had participated but no one fully understood the outcome.
Beside him the augmentation system updated its recommendation weighting.
LIKELIHOOD OF GOVERNANCE NEGLIGENCE: 87%
The percentage shifted upward slightly as further analysis completed. Judge Pharos looked at the number for a moment longer than he intended.
Then below the percentage another field appeared silently on the display.
BOARD ACCOUNTABILITY ATTRIBUTION: CONTRIBUTORY NEGLIGENCE LIKELY
Judge Pharos stared at the wording for a moment.
Twenty years ago the law had focused on intent. Now the systems focused on influence, oversight, delegated authority and failure to intervene. The distinction had changed everything and placed a huge amount of interpretive power into the hands of AI.
Now that AI had carved out realms of subtle authority within society, it was hard to resist its influence. Not the technology itself, but the quiet psychological gravity of its guidance. The machine never forced a decision, it simply presented probabilities, patterns and recommendations. Yet over time, fewer and fewer people disagreed with it.
Not because it was always right, but because it was usually right, and eventually society stopped questioning the difference.
Judge Pharos looked back toward Phillip Miles sitting alone in the witness enclosure.
The court would eventually determine whether the board had acted negligently.
But sitting there in silence, watching the live recommendations continue to update beside him, another question lingered quietly in the back of his mind.
At what point had human judgement itself become little more than a rubber stamp for machine certainty?
(Human Written First – AI Editied Second)
